fix(llm): Ensure proper JSON encoding by setting ensure_ascii=False in tool response serialization

backend/onyx/llm/utils.py

@@ -345,7 +345,7 @@ def check_message_tokens(
     if isinstance(message, AIMessage) and message.tool_calls:
         for tool_call in message.tool_calls:
             total_tokens += check_number_of_tokens(
-                json.dumps(tool_call["args"]), encode_fn
+                json.dumps(tool_call["args"], ensure_ascii=False), encode_fn
             )
             total_tokens += check_number_of_tokens(tool_call["name"], encode_fn)
 

backend/onyx/tools/tool_implementations/custom/custom_tool.py

@@ -140,7 +140,7 @@ class CustomTool(BaseTool):
             return json.dumps({"file_ids": image_response.file_ids})
 
         # For JSON or other responses, return as-is
-        return json.dumps(response.tool_result)
+        return json.dumps(response.tool_result, ensure_ascii=False)
 
     """For LLMs which do NOT support explicit tool calling"""
 

backend/onyx/tools/tool_implementations/images/image_generation_tool.py

@@ -236,7 +236,8 @@ class ImageGenerationTool(Tool[None]):
                         "url": image_generation.url,
                     }
                     for image_generation in image_generations
-                ]
+                ],
+                ensure_ascii=False
             ),
         )
 

backend/onyx/tools/tool_implementations/mcp/mcp_tool.py

@@ -223,7 +223,7 @@ Return ONLY a valid JSON object with the extracted arguments. If no arguments ar
                 response=CustomToolCallSummary(
                     tool_name=self._name,
                     response_type="json",
-                    tool_result=json.dumps({"tool_result": tool_result}),
+                    tool_result=json.dumps({"tool_result": tool_result}, ensure_ascii=False),
                 ),
             )
 

backend/onyx/tools/tool_implementations/search/search_tool.py

@@ -225,7 +225,8 @@ class SearchTool(Tool[SearchToolOverrideKwargs]):
                     llm_doc_to_dict(doc, ind)
                     for ind, doc in enumerate(final_context_docs)
                 ]
-            }
+            },
+            ensure_ascii=False
         )
 
     """For LLMs that don't support tool calling"""